Munoz and Burton: Video observations of sharks in reef habitats off east-central Florida 
89 
distance between 2 traps was 237 m, but these 2 sites 
were sampled more than 3 h apart. 
Underwater video 
We outfitted each chevron trap with 2 outward-looking 
video cameras, forming a 2-camera BRUVS. A high- 
definition Vixia HF S200 2 video camera (Canon U.S.A., 
Inc., Melville, NY) with a standard lens in an HF-S21 
underwater housing (Gates Underwater Products, Inc., 
Poway, CA) was attached over the mouth of each trap, 
facing away from the trap. A second high-definition vid¬ 
2 Mention of trade names or commercial companies is for iden¬ 
tification purposes only and does not imply endorsement by 
the National Marine Fisheries Service, NOAA. 
eo camera (HD HER04 Black, GoPro, Inc., San 
Mateo, CA) was attached over the nose of each 
trap (opposite the Canon camera), also looking 
outward. The Canon camera was used to re¬ 
cord video for quantifying the number of fish 
at each site (i.e., for video analysis, a single, 
unique Canon video recording was reviewed 
for each site). The GoPro camera, which pro¬ 
vided additional and sometimes better views 
of activity at a site if a fish passed closer to 
it, was used only when necessary to confirm 
identification of fish species. 
Video recordings were ~90 min in length. 
The SERFS video-analysis protocol commences 
10 min after a trap reaches the seafloor, with 
fish counted every 30 s during a 20-min period 
(or segment) for a total of 41 frames, each 1 s 
in duration. For this study, we reviewed videos 
in their entirety (rather than a subset of forty- 
one 1-s frames) at a speed that was approxi¬ 
mately 2 times faster than the viewing speed 
used in the standard protocol, using a jog 
wheel (as part of our video editing hardware 
setup) for precise control of video speed. The 
jog wheel allowed us to rapidly review videos 
without exceeding a maximum rate that would 
have resulted in skipped video frames and po¬ 
tentially in missed fish detections. All sharks 
observed were identified (enabling collection 
of presence-absence data for each species and 
video recording) and counted, and, when pos¬ 
sible, their sex was determined to be male if 
claspers were observed (sex data not shown). 
We compared the trap catch between sites with 
and without video observations of sharks by 
using a Mann-Whitney U test. We calculated 
frequency of occurrence for each species as the 
number of sites with observations of a given 
shark species divided by the total number of 
sampled sites. 
We compared, using a chi-square test, the 
proportion of sites with sharks observed dur¬ 
ing the SERFS 20-min segment with the pro¬ 
portion of sites with sharks observed outside 
the 20-min segment (i.e., before, or the first 10 min 
after the trap reaches the seafloor, and after, or the 
part of the video from 30 min after the trap reaches 
the bottom to the end). Videos with observations of 
sharks were classified as either outside or within the 
20-min segment. Videos classified as outside the 20- 
min segment had shark observations solely outside the 
segment. Videos classified as within the 20-min seg¬ 
ment could include observations outside the segment in 
addition to observations within the segment, but the 2 
classifications were mutually exclusive. 
Underwater visual census 
Within 10 min of the deployment of each BRUVS, 2 
divers descended along the trap line until they could 
